- 23 Feb, 2017 15 commits
-
-
Yinghai Lu authored
commit afe3e4d1 upstream. In addition to making PME non-modular, d7def204 ("PCI/PME: Make explicitly non-modular") removed the pcie_pme_driver .remove() method, pcie_pme_remove(). pcie_pme_remove() freed the PME IRQ that was requested in pci_pme_probe(). The fact that we don't free the IRQ after d7def204 causes the following crash when removing a PCIe port device via /sys: ------------[ cut here ]------------ kernel BUG at drivers/pci/msi.c:370! invalid opcode: 0000 [#1] SMP Modules linked in: CPU: 1 PID: 14509 Comm: sh Tainted: G W 4.8.0-rc1-yh-00012-gd29438d6 RIP: 0010:[<ffffffff9758bbf5>] free_msi_irqs+0x65/0x190 ... Call Trace: [<ffffffff9758cda4>] pci_disable_msi+0x34/0x40 [<ffffffff97583817>] cleanup_service_irqs+0x27/0x30 [<ffffffff97583e9a>] pcie_port_device_remove+0x2a/0x40 [<ffffffff97584250>] pcie_portdrv_remove+0x40/0x50 [<ffffffff97576d7b>] pci_device_remove+0x4b/0xc0 [<ffffffff9785ebe6>] __device_release_driver+0xb6/0x150 [<ffffffff9785eca5>] device_release_driver+0x25/0x40 [<ffffffff975702e4>] pci_stop_bus_device+0x74/0xa0 [<ffffffff975704ea>] pci_stop_and_remove_bus_device_locked+0x1a/0x30 [<ffffffff97578810>] remove_store+0x50/0x70 [<ffffffff9785a378>] dev_attr_store+0x18/0x30 [<ffffffff97260b64>] sysfs_kf_write+0x44/0x60 [<ffffffff9725feae>] kernfs_fop_write+0x10e/0x190 [<ffffffff971e13f8>] __vfs_write+0x28/0x110 [<ffffffff970b0fa4>] ? percpu_down_read+0x44/0x80 [<ffffffff971e53a7>] ? __sb_start_write+0xa7/0xe0 [<ffffffff971e53a7>] ? __sb_start_write+0xa7/0xe0 [<ffffffff971e1f04>] vfs_write+0xc4/0x180 [<ffffffff971e3089>] SyS_write+0x49/0xa0 [<ffffffff97001a46>] do_syscall_64+0xa6/0x1b0 [<ffffffff9819201e>] entry_SYSCALL64_slow_path+0x25/0x25 ... RIP [<ffffffff9758bbf5>] free_msi_irqs+0x65/0x190 RSP <ffff89ad3085bc48> ---[ end trace f4505e1dac5b95d3 ]--- Segmentation fault Restore pcie_pme_remove(). [bhelgaas: changelog] Fixes: d7def204 ("PCI/PME: Make explicitly non-modular") Signed-off-by:
Yinghai Lu <yinghai@kernel.org> Signed-off-by:
Bjorn Helgaas <bhelgaas@google.com> Acked-by:
Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
Jarkko Nikula authored
commit 12688dc2 upstream. This reverts commit 63d0f0a6. It caused a regression on platforms where I2C controller is synthesized with dynamic TAR update disabled. Detection code is testing is bit DW_IC_CON_10BITADDR_MASTER in register DW_IC_CON read-only but fails to restore original value in case bit is read-write. Instead of fixing this we revert the commit since it was preparation for the commit 0317e6c0 ("i2c: designware: do not disable adapter after transfer") which was also reverted. Reported-by:
Shah Nehal-Bakulchandra <Nehal-bakulchandra.Shah@amd.com> Reported-by:
Suravee Suthikulpanit <suravee.suthikulpanit@amd.com> Acked-By:
Lucas De Marchi <lucas.demarchi@intel.com> Fixes: 63d0f0a6 ("i2c: designware: detect when dynamic tar update is possible") Signed-off-by:
Jarkko Nikula <jarkko.nikula@linux.intel.com> Signed-off-by:
Wolfram Sang <wsa@the-dreams.de> Signed-off-by:
-
Kees Cook authored
commit 9e344048 upstream. The 64-bit get_user() wasn't clearing the high word due to a typo in the error handler. The exception handler entry was already correct, though. Noticed during recent usercopy test additions in lib/test_user_copy.c. Signed-off-by:
Kees Cook <keescook@chromium.org> Signed-off-by:
Russell King <rmk+kernel@armlinux.org.uk> Signed-off-by:
-
Yang Yang authored
commit 25f71d1c upstream. The UEVENT user mode helper is enabled before the initcalls are executed and is available when the root filesystem has been mounted. The user mode helper is triggered by device init calls and the executable might use the futex syscall. futex_init() is marked __initcall which maps to device_initcall, but there is no guarantee that futex_init() is invoked _before_ the first device init call which triggers the UEVENT user mode helper. If the user mode helper uses the futex syscall before futex_init() then the syscall crashes with a NULL pointer dereference because the futex subsystem has not been initialized yet. Move futex_init() to core_initcall so futexes are initialized before the root filesystem is mounted and the usermode helper becomes available. [ tglx: Rewrote changelog ] Signed-off-by:
Yang Yang <yang.yang29@zte.com.cn> Cc: jiang.biao2@zte.com.cn Cc: jiang.zhengxiong@zte.com.cn Cc: zhong.weidong@zte.com.cn Cc: deng.huali@zte.com.cn Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/1483085875-6130-1-git-send-email-yang.yang29@zte.com.cnSigned-off-by:
Thomas Gleixner <tglx@linutronix.de> Signed-off-by:
-
Pierre-Louis Bossart authored
commit bb08c04d upstream. 100% reproducible issue found on SKL SkullCanyon NUC with two external DP daisy-chained monitors in DP/MST mode. When turning off or changing the input of the second monitor the machine stops with a kernel oops. This issue happened with 4.8.8 as well as drm/drm-intel-nightly. This issue is traced to an inconsistent control flow in drm_dp_update_payload_part1(): the 'port' pointer is set to NULL at the same time as 'req_payload.num_slots' is set to zero, but the pointer is dereferenced even when req_payload.num_slot is zero. The problematic dereference was introduced in commit dfda0df3 ("drm/mst: rework payload table allocation to conform better") and may impact all versions since v3.18 The fix suggested by Chris Wilson removes the kernel oops and was found to work well after 10mn of monkey-testing with the second monitor power and input buttons Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98990 Fixes: dfda0df3 ("drm/mst: rework payload table allocation to conform better.") Cc: Dave Airlie <airlied@redhat.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Nathan D Ciobanu <nathan.d.ciobanu@linux.intel.com> Cc: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com> Cc: Sean Paul <seanpaul@chromium.org> Tested-by:
Nathan D Ciobanu <nathan.d.ciobanu@linux.intel.com> Reviewed-by:
Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com> Signed-off-by:
Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Signed-off-by:
Jani Nikula <jani.nikula@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1487076561-2169-1-git-send-email-jani.nikula@intel.comSigned-off-by:
-
Michel Dänzer authored
commit d74c67dd upstream. The crtc_h/vdisplay fields may not match the CRTC viewport dimensions with special modes such as interlaced ones. Fixes the HW cursor disappearing in the bottom half of the screen with interlaced modes. Fixes: 6b16cf77 ("drm/radeon: Hide the HW cursor while it's out of bounds") Reported-by:
Ashutosh Kumar <ashutosh.kumar@amd.com> Tested-by:
Sonny Jiang <sonny.jiang@amd.com> Reviewed-by:
Alex Deucher <alexander.deucher@amd.com> Signed-off-by:
Michel Dänzer <michel.daenzer@amd.com> Signed-off-by:
-
IHARA Hiroka authored
commit 722c5ac7 upstream. ELAN0605 has been confirmed to be a variant of ELAN0600, which is blacklisted in the hid-core to be managed by elan_i2c. This device can be found in Lenovo ideapad 310s (80U4000). Signed-off-by:
Hiroka IHARA <ihara_h@live.jp> Signed-off-by:
Dmitry Torokhov <dmitry.torokhov@gmail.com> Signed-off-by:
-
Al Viro authored
commit 137d01df upstream. What happens is that a write to /dev/sg is given a request with non-zero ->iovec_count combined with zero ->dxfer_len. Or with ->dxferp pointing to an array full of empty iovecs. Having write permission to /dev/sg shouldn't be equivalent to the ability to trigger BUG_ON() while holding spinlocks... Found by Dmitry Vyukov and syzkaller. [ The BUG_ON() got changed to a WARN_ON_ONCE(), but this fixes the underlying issue. - Linus ] Signed-off-by:
Al Viro <viro@zeniv.linux.org.uk> Reported-by:
Dmitry Vyukov <dvyukov@google.com> Reviewed-by:
Christoph Hellwig <hch@lst.de> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by:
-
Johannes Thumshirn authored
commit fd3fc0b4 upstream. Don't crash the machine just because of an empty transfer. Use WARN_ON() combined with returning an error. Found by Dmitry Vyukov and syzkaller. [ Changed to "WARN_ON_ONCE()". Al has a patch that should fix the root cause, but a BUG_ON() is not acceptable in any case, and a WARN_ON() might still be a cause of excessive log spamming. NOTE! If this warning ever triggers, we may end up leaking resources, since this doesn't bother to try to clean the command up. So this WARN_ON_ONCE() triggering does imply real problems. But BUG_ON() is much worse. People really need to stop using BUG_ON() for "this shouldn't ever happen". It makes pretty much any bug worse. - Linus ] Signed-off-by:
Johannes Thumshirn <jthumshirn@suse.de> Reported-by:
-
Paul Mackerras authored
commit 3f91a89d upstream. Currently, if the kernel is running on a POWER9 processor under a hypervisor, it may try to use the radix MMU even though it doesn't have the necessary code to do so (it doesn't negotiate use of radix, and it doesn't do the H_REGISTER_PROC_TBL hcall). If the hypervisor supports both radix and HPT, then it will set up the guest to use HPT (since the guest doesn't request radix in the CAS call), but if the radix feature bit is set in the ibm,pa-features property (which is valid, since ibm,pa-features is defined to represent the capabilities of the processor) the guest will try to use radix, resulting in a crash when it turns the MMU on. This makes the minimal fix for the current code, which is to disable radix unless we are running in hypervisor mode. Fixes: 2bfd65e4 ("powerpc/mm/radix: Add radix callbacks for early init routines") Signed-off-by:
Paul Mackerras <paulus@ozlabs.org> Signed-off-by:
Michael Ellerman <mpe@ellerman.id.au> Signed-off-by:
-
Anssi Hannula authored
commit 3d4ef329 upstream. Commit 577fb131 ("mmc: rework selection of bus speed mode") refactored bus width selection code to mmc_select_bus_width(). However, it also altered the behavior to not call the selection code in non-high-speed modes anymore. This causes 1-bit mode to always be used when the high-speed mode is not enabled, even though 4-bit and 8-bit bus are valid bus widths in the backwards-compatibility (legacy) mode as well (see e.g. 5.3.2 Bus Speed Modes in JEDEC 84-B50). This results in a significant regression in transfer speeds. Fix the code to allow 4-bit and 8-bit widths even without high-speed mode, as before. Tested with a Zynq-7000 PicoZed 7020 board. Fixes: 577fb131 ("mmc: rework selection of bus speed mode") Signed-off-by:
Anssi Hannula <anssi.hannula@bitwise.fi> Signed-off-by:
Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by:
-
Miklos Szeredi authored
commit 84588a93 upstream. Signed-off-by:
Miklos Szeredi <mszeredi@redhat.com> Fixes: d82718e3 ("fuse_dev_splice_read(): switch to add_to_pipe()") Signed-off-by:
-
Sahitya Tummala authored
commit 6ba4d272 upstream. There is a potential race between fuse_dev_do_write() and request_wait_answer() contexts as shown below: TASK 1: __fuse_request_send(): |--spin_lock(&fiq->waitq.lock); |--queue_request(); |--spin_unlock(&fiq->waitq.lock); |--request_wait_answer(): |--if (test_bit(FR_SENT, &req->flags)) <gets pre-empted after it is validated true> TASK 2: fuse_dev_do_write(): |--clears bit FR_SENT, |--request_end(): |--sets bit FR_FINISHED |--spin_lock(&fiq->waitq.lock); |--list_del_init(&req->intr_entry); |--spin_unlock(&fiq->waitq.lock); |--fuse_put_request(); |--queue_interrupt(); <request gets queued to interrupts list> |--wake_up_locked(&fiq->waitq); |--wait_event_freezable(); <as FR_FINISHED is set, it returns and then the caller frees this request> Now, the next fuse_dev_do_read(), see interrupts list is not empty and then calls fuse_read_interrupt() which tries to access the request which is already free'd and gets the below crash: [11432.401266] Unable to handle kernel paging request at virtual address 6b6b6b6b6b6b6b6b ... [11432.418518] Kernel BUG at ffffff80083720e0 [11432.456168] PC is at __list_del_entry+0x6c/0xc4 [11432.463573] LR is at fuse_dev_do_read+0x1ac/0x474 ... [11432.679999] [<ffffff80083720e0>] __list_del_entry+0x6c/0xc4 [11432.687794] [<ffffff80082c65e0>] fuse_dev_do_read+0x1ac/0x474 [11432.693180] [<ffffff80082c6b14>] fuse_dev_read+0x6c/0x78 [11432.699082] [<ffffff80081d5638>] __vfs_read+0xc0/0xe8 [11432.704459] [<ffffff80081d5efc>] vfs_read+0x90/0x108 [11432.709406] [<ffffff80081d67f0>] SyS_read+0x58/0x94 As FR_FINISHED bit is set before deleting the intr_entry with input queue lock in request completion path, do the testing of this flag and queueing atomically with the same lock in queue_interrupt(). Signed-off-by:
Sahitya Tummala <stummala@codeaurora.org> Signed-off-by:
-
Mauro Carvalho Chehab authored
commit f9c85ee6 upstream. Reported as a Kaffeine bug: https://bugs.kde.org/show_bug.cgi?id=375811 The USB control messages require DMA to work. We cannot pass a stack-allocated buffer, as it is not warranted that the stack would be into a DMA enabled area. On Kernel 4.9, the default is to not accept DMA on stack anymore on x86 architecture. On other architectures, this has been a requirement since Kernel 2.2. So, after this patch, this driver should likely work fine on all archs. Tested with USB ID 2040:5510: Hauppauge Windham Signed-off-by:
Mauro Carvalho Chehab <mchehab@s-opensource.com> Signed-off-by:
-
Miklos Szeredi authored
commit 5a81e6a1 upstream. Flags (PIPE_BUF_FLAG_PACKET, PIPE_BUF_FLAG_GIFT) could remain on the unused part of the pipe ring buffer. Previously splice_to_pipe() left the flags value alone, which could result in incorrect behavior. Uninitialized flags appears to have been there from the introduction of the splice syscall. Signed-off-by:
-
- 18 Feb, 2017 25 commits
-
-
Greg Kroah-Hartman authored
-
Yu-cheng Yu authored
commit dffba9a3 upstream. The compacted-format XSAVES area is determined at boot time and never changed after. The field xsave.header.xcomp_bv indicates which components are in the fixed XSAVES format. In fpstate_init() we did not set xcomp_bv to reflect the XSAVES format since at the time there is no valid data. However, after we do copy_init_fpstate_to_fpregs() in fpu__clear(), as in commit: b22cbe40 x86/fpu: Fix invalid FPU ptrace state after execve() and when __fpu_restore_sig() does fpu__restore() for a COMPAT-mode app, a #GP occurs. This can be easily triggered by doing valgrind on a COMPAT-mode "Hello World," as reported by Joakim Tjernlund and others: https://bugzilla.kernel.org/show_bug.cgi?id=190061 Fix it by setting xcomp_bv correctly. This patch also moves the xcomp_bv initialization to the proper place, which was in copyin_to_xsaves() as of: 4c833368 x86/fpu: Set the xcomp_bv when we fake up a XSAVES area which fixed the bug too, but it's more efficient and cleaner to initialize things once per boot, not for every signal handling operation. Reported-by:
Kevin Hao <haokexin@gmail.com> Reported-by:
Joakim Tjernlund <Joakim.Tjernlund@infinera.com> Signed-off-by:
Yu-cheng Yu <yu-cheng.yu@intel.com> Cc: Andy Lutomirski <luto@kernel.org> Cc: Borislav Petkov <bp@suse.de> Cc: Dave Hansen <dave.hansen@linux.intel.com> Cc: Fenghua Yu <fenghua.yu@intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ravi V. Shankar <ravi.v.shankar@intel.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: haokexin@gmail.com Link: http://lkml.kernel.org/r/1485212084-4418-1-git-send-email-yu-cheng.yu@intel.com [ Combined it with 4c833368. ] Signed-off-by:
Ingo Molnar <mingo@kernel.org> Signed-off-by:
-
Pablo Neira authored
commit 92e55f41 upstream. Unlike ipv4, this control socket is shared by all cpus so we cannot use it as scratchpad area to annotate the mark that we pass to ip6_xmit(). Add a new parameter to ip6_xmit() to indicate the mark. The SCTP socket family caches the flowi6 structure in the sctp_transport structure, so we cannot use to carry the mark unless we later on reset it back, which I discarded since it looks ugly to me. Fixes: bf99b4de ("tcp: fix mark propagation with fwmark_reflect enabled") Suggested-by:
Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by:
Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by:
David S. Miller <davem@davemloft.net> Signed-off-by:
-
Mark Bloch authored
commit 0fd758d6 upstream. When adding a new rule to an fte, we need to hold the fte lock until we add that rule to the fte and increase the fte ref count. Fixes: 0c56b975 ("net/mlx5_core: Introduce flow steering API") Signed-off-by:
Mark Bloch <markb@mellanox.com> Signed-off-by:
Saeed Mahameed <saeedm@mellanox.com> Signed-off-by:
Leon Romanovsky <leon@kernel.org> Signed-off-by:
-
Pau Espin Pedrol authored
commit bf99b4de upstream. Otherwise, RST packets generated by the TCP stack for non-existing sockets always have mark 0. The mark from the original packet is assigned to the netns_ipv4/6 socket used to send the response so that it can get copied into the response skb when the socket sends it. Fixes: e110861f ("net: add a sysctl to reflect the fwmark on replies") Cc: Lorenzo Colitti <lorenzo@google.com> Signed-off-by:
Pau Espin Pedrol <pau.espin@tessares.net> Signed-off-by:
-
Hangbin Liu authored
[ Upstream commit 9c8bb163 ] In function igmpv3/mld_add_delrec() we allocate pmc and put it in idev->mc_tomb, so we should free it when we don't need it in del_delrec(). But I removed kfree(pmc) incorrectly in latest two patches. Now fix it. Fixes: 24803f38 ("igmp: do not remove igmp souce list info when ...") Fixes: 1666d49e ("mld: do not remove mld souce list info when ...") Reported-by:
Daniel Borkmann <daniel@iogearbox.net> Signed-off-by:
Hangbin Liu <liuhangbin@gmail.com> Signed-off-by:
-
Hangbin Liu authored
[ Upstream commit 1666d49e ] This is an IPv6 version of commit 24803f38 ("igmp: do not remove igmp souce list..."). In mld_del_delrec(), we will restore back all source filter info instead of flush them. Move mld_clear_delrec() from ipv6_mc_down() to ipv6_mc_destroy_dev() since we should not remove source list info when set link down. Remove igmp6_group_dropped() in ipv6_mc_destroy_dev() since we have called it in ipv6_mc_down(). Also clear all source info after igmp6_group_dropped() instead of in it because ipv6_mc_down() will call igmp6_group_dropped(). Signed-off-by:
-
Eric Dumazet authored
[ Upstream commit 72fb96e7 ] udp_ioctl(), as its name suggests, is used by UDP protocols, but is also used by L2TP :( L2TP should use its own handler, because it really does not look the same. SIOCINQ for instance should not assume UDP checksum or headers. Thanks to Andrey and syzkaller team for providing the report and a nice reproducer. While crashes only happen on recent kernels (after commit 7c13f97f ("udp: do fwd memory scheduling on dequeue")), this probably needs to be backported to older kernels. Fixes: 7c13f97f ("udp: do fwd memory scheduling on dequeue") Fixes: 85584672 ("udp: Fix udp_poll() and ioctl()") Signed-off-by:
Eric Dumazet <edumazet@google.com> Reported-by:
Andrey Konovalov <andreyknvl@google.com> Acked-by:
Paolo Abeni <pabeni@redhat.com> Signed-off-by:
-
Florian Fainelli authored
[ Upstream commit 382e1eea ] dsa_slave_create() can fail, and dsa_user_port_unapply() will properly check for the network device not being NULL before attempting to destroy it. We were not setting the slave network device as NULL if dsa_slave_create() failed, so we would later on be calling dsa_slave_destroy() on a now free'd and unitialized network device, causing crashes in dsa_slave_destroy(). Fixes: 83c0afae ("net: dsa: Add new binding implementation") Signed-off-by:
Florian Fainelli <f.fainelli@gmail.com> Signed-off-by:
-
WANG Cong authored
[ Upstream commit 73d2c667 ] Andrey reported a kernel crash: general protection fault: 0000 [#1] SMP KASAN Dumping ftrace buffer: (ftrace buffer empty) Modules linked in: CPU: 2 PID: 3880 Comm: syz-executor1 Not tainted 4.10.0-rc6+ #124 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011 task: ffff880060048040 task.stack: ffff880069be8000 RIP: 0010:ping_v4_push_pending_frames net/ipv4/ping.c:647 [inline] RIP: 0010:ping_v4_sendmsg+0x1acd/0x23f0 net/ipv4/ping.c:837 RSP: 0018:ffff880069bef8b8 EFLAGS: 00010206 RAX: dffffc0000000000 RBX: ffff880069befb90 RCX: 0000000000000000 RDX: 0000000000000018 RSI: ffff880069befa30 RDI: 00000000000000c2 RBP: ffff880069befbb8 R08: 0000000000000008 R09: 0000000000000000 R10: 0000000000000002 R11: 0000000000000000 R12: ffff880069befab0 R13: ffff88006c624a80 R14: ffff880069befa70 R15: 0000000000000000 FS: 00007f6f7c716700(0000) GS:ffff88006de00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00000000004a6f28 CR3: 000000003a134000 CR4: 00000000000006e0 Call Trace: inet_sendmsg+0x164/0x5b0 net/ipv4/af_inet.c:744 sock_sendmsg_nosec net/socket.c:635 [inline] sock_sendmsg+0xca/0x110 net/socket.c:645 SYSC_sendto+0x660/0x810 net/socket.c:1687 SyS_sendto+0x40/0x50 net/socket.c:1655 entry_SYSCALL_64_fastpath+0x1f/0xc2 This is because we miss a check for NULL pointer for skb_peek() when the queue is empty. Other places already have the same check. Fixes: c319b4d7 ("net: ipv4: add IPPROTO_ICMP socket kind") Reported-by:
Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by:
-
Willem de Bruijn authored
[ Upstream commit 57031eb7 ] Link layer protocols may unconditionally pull headers, as Ethernet does in eth_type_trans. Ensure that the entire link layer header always lies in the skb linear segment. tpacket_snd has such a check. Extend this to packet_snd. Variable length link layer headers complicate the computation somewhat. Here skb->len may be smaller than dev->hard_header_len. Round up the linear length to be at least as long as the smallest of the two. Reported-by:
Willem de Bruijn <willemb@google.com> Acked-by:
-
Willem de Bruijn authored
[ Upstream commit 217e6fa2 ] The stack must not pass packets to device drivers that are shorter than the minimum link layer header length. Previously, packet sockets would drop packets smaller than or equal to dev->hard_header_len, but this has false positives. Zero length payload is used over Ethernet. Other link layer protocols support variable length headers. Support for validation of these protocols removed the min length check for all protocols. Introduce an explicit dev->min_header_len parameter and drop all packets below this value. Initially, set it to non-zero only for Ethernet and loopback. Other protocols can follow in a patch to net-next. Fixes: 9ed988cd ("packet: validate variable length ll headers") Reported-by:
Sowmini Varadhan <sowmini.varadhan@oracle.com> Signed-off-by:
-
WANG Cong authored
[ Upstream commit d7426c69 ] Dmitry reported a double free in sit_init_net(): kernel BUG at mm/percpu.c:689! invalid opcode: 0000 [#1] SMP KASAN Dumping ftrace buffer: (ftrace buffer empty) Modules linked in: CPU: 0 PID: 15692 Comm: syz-executor1 Not tainted 4.10.0-rc6-next-20170206 #1 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 task: ffff8801c9cc27c0 task.stack: ffff88017d1d8000 RIP: 0010:pcpu_free_area+0x68b/0x810 mm/percpu.c:689 RSP: 0018:ffff88017d1df488 EFLAGS: 00010046 RAX: 0000000000010000 RBX: 00000000000007c0 RCX: ffffc90002829000 RDX: 0000000000010000 RSI: ffffffff81940efb RDI: ffff8801db841d94 RBP: ffff88017d1df590 R08: dffffc0000000000 R09: 1ffffffff0bb3bdd R10: dffffc0000000000 R11: 00000000000135dd R12: ffff8801db841d80 R13: 0000000000038e40 R14: 00000000000007c0 R15: 00000000000007c0 FS: 00007f6ea608f700(0000) GS:ffff8801dbe00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 000000002000aff8 CR3: 00000001c8d44000 CR4: 00000000001426f0 DR0: 0000000020000000 DR1: 0000000020000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000600 Call Trace: free_percpu+0x212/0x520 mm/percpu.c:1264 ipip6_dev_free+0x43/0x60 net/ipv6/sit.c:1335 sit_init_net+0x3cb/0xa10 net/ipv6/sit.c:1831 ops_init+0x10a/0x530 net/core/net_namespace.c:115 setup_net+0x2ed/0x690 net/core/net_namespace.c:291 copy_net_ns+0x26c/0x530 net/core/net_namespace.c:396 create_new_namespaces+0x409/0x860 kernel/nsproxy.c:106 unshare_nsproxy_namespaces+0xae/0x1e0 kernel/nsproxy.c:205 SYSC_unshare kernel/fork.c:2281 [inline] SyS_unshare+0x64e/0xfc0 kernel/fork.c:2231 entry_SYSCALL_64_fastpath+0x1f/0xc2 This is because when tunnel->dst_cache init fails, we free dev->tstats once in ipip6_tunnel_init() and twice in sit_init_net(). This looks redundant but its ndo_uinit() does not seem enough to clean up everything here. So avoid this by setting dev->tstats to NULL after the first free, at least for -net. Reported-by:
-
David Ahern authored
[ Upstream commit 2bd137de ] An error was reported upgrading to 4.9.8: root@Typhoon:~# ip route add default table 210 nexthop dev eth0 via 10.68.64.1 weight 1 nexthop dev eth0 via 10.68.64.2 weight 1 RTNETLINK answers: Operation not supported The problem occurs when CONFIG_LWTUNNEL is not enabled and a multipath route is submitted. The point of lwtunnel_valid_encap_type_attr is catch modules that need to be loaded before any references are taken with rntl held. With CONFIG_LWTUNNEL disabled, there will be no modules to load so the lwtunnel_valid_encap_type_attr stub should just return 0. Fixes: 9ed59592 ("lwtunnel: fix autoload of lwt modules") Reported-by: pupilla@libero.it Signed-off-by:
David Ahern <dsa@cumulusnetworks.com> Signed-off-by:
-
Marcelo Ricardo Leitner authored
[ Upstream commit 2dcab598 ] Alexander Popov reported that an application may trigger a BUG_ON in sctp_wait_for_sndbuf if the socket tx buffer is full, a thread is waiting on it to queue more data and meanwhile another thread peels off the association being used by the first thread. This patch replaces the BUG_ON call with a proper error handling. It will return -EPIPE to the original sendmsg call, similarly to what would have been done if the association wasn't found in the first place. Acked-by:
Alexander Popov <alex.popov@linux.com> Signed-off-by:
Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> Reviewed-by:
Xin Long <lucien.xin@gmail.com> Signed-off-by:
-
Benjamin Poirier authored
[ Upstream commit bd4ce941 ] mlx4 may schedule napi from a workqueue. Afterwards, softirqs are not run in a deterministic time frame and the following message may be logged: NOHZ: local_softirq_pending 08 The problem is the same as what was described in commit ec13ee80 ("virtio_net: invoke softirqs after __napi_schedule") and this patch applies the same fix to mlx4. Fixes: 07841f9d ("net/mlx4_en: Schedule napi when RX buffers allocation fails") Cc: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by:
Benjamin Poirier <bpoirier@suse.com> Acked-by:
Tariq Toukan <tariqt@mellanox.com> Signed-off-by:
-
Ben Hutchings authored
[ Upstream commit 2d6a0e9d ] Allocating USB buffers on the stack is not portable, and no longer works on x86_64 (with VMAP_STACK enabled as per default). Fixes: 1da177e4 ("Linux-2.6.12-rc2") Signed-off-by:
Ben Hutchings <ben@decadent.org.uk> Signed-off-by:
-
Ben Hutchings authored
[ Upstream commit d4114914 ] Signed-off-by:
-
Ben Hutchings authored
[ Upstream commit 7926aff5 ] Allocating USB buffers on the stack is not portable, and no longer works on x86_64 (with VMAP_STACK enabled as per default). Fixes: 1da177e4 ("Linux-2.6.12-rc2") Signed-off-by:
-
Ben Hutchings authored
[ Upstream commit 5593523f ] Allocating USB buffers on the stack is not portable, and no longer works on x86_64 (with VMAP_STACK enabled as per default). Fixes: 1da177e4 ("Linux-2.6.12-rc2") References: https://bugs.debian.org/852556Reported-by:
Lisandro Damián Nicanor Pérez Meyer <lisandro@debian.org> Tested-by:
-
Willem de Bruijn authored
[ Upstream commit 837585a5 ] When IFF_VNET_HDR is enabled, a virtio_net header must precede data. Data length is verified to be greater than or equal to expected header length tun->vnet_hdr_sz before copying. Macvtap functions read the value once, but unless READ_ONCE is used, the compiler may ignore this and read multiple times. Enforce a single read and locally cached value to avoid updates between test and use. Signed-off-by:
-
Willem de Bruijn authored
[ Upstream commit e1edab87 ] When IFF_VNET_HDR is enabled, a virtio_net header must precede data. Data length is verified to be greater than or equal to expected header length tun->vnet_hdr_sz before copying. Read this value once and cache locally, as it can be updated between the test and use (TOCTOU). Signed-off-by:
-
Eric Dumazet authored
[ Upstream commit ccf7abb9 ] Splicing from TCP socket is vulnerable when a packet with URG flag is received and stored into receive queue. __tcp_splice_read() returns 0, and sk_wait_data() immediately returns since there is the problematic skb in queue. This is a nice way to burn cpu (aka infinite loop) and trigger soft lockups. Again, this gem was found by syzkaller tool. Fixes: 9c55e01c ("[TCP]: Splice receive support.") Signed-off-by:
-
Eric Dumazet authored
[ Upstream commit ebf6c9cb ] Dmitry reported use-after-free in ip6_datagram_recv_specific_ctl() A similar bug was fixed in commit 8ce48623 ("ipv6: tcp: restore IP6CB for pktoptions skbs"), but I missed another spot. tcp_v6_syn_recv_sock() can indeed set np->pktoptions from ireq->pktopts Fixes: 971f10ec ("tcp: better TCP_SKB_CB layout to reduce cache line misses") Signed-off-by:
-
Eric Dumazet authored
[ Upstream commit 7892032c ] Andrey Konovalov reported out of bound accesses in ip6gre_err() If GRE flags contains GRE_KEY, the following expression *(((__be32 *)p) + (grehlen / 4) - 1) accesses data ~40 bytes after the expected point, since grehlen includes the size of IPv6 headers. Let's use a "struct gre_base_hdr *greh" pointer to make this code more readable. p[1] becomes greh->protocol. grhlen is the GRE header length. Fixes: c12b395a ("gre: Support GRE over IPv6") Signed-off-by:
-
